ROCK CUTTING FORCE ESTIMATION IN TUNNELING WITH TBM

Abstract

The performance of TBM affects the execution cost and completion time of the rock excavation project. Therefore, it is vital to correctly predict the performance of TBM. Despite the large amount of research about TBM performance and estimation of its parameters, there is still a gap. Predicting cutting force remains a complex task due to the variability in rock conditions and properties, the diversity of TBM types, and the need to consider all relevant parameters and properties together. Therefore, it is necessary to analyze data using regression models based on statistical analysis. This thesis aims to address the performance prediction problem and improve the performance prediction model by gaining a better understanding of the interaction between rock and cutting force. To achieve the goal of this study, simple linear, multilinear and nonlinear regression analysis based on statistical analysis approaches were employed to develop a series of TBM performance model. A comprehensive database of TBM performance compiled from 3 tunnelling projects of Iran (Zagros, Ghomrood and Karaj), was established and used for the development of the model. The results of the study showed the influence of different rock parameters on the cutting force. Also, the quality of the rock has a significant impact on the cutting force. The results indicated that non-linear equations are more robust than linear models because linear relationships are less realistic under such volatile and unpredictable conditions. Compared to previous research, the current model, which utilizes intact rock properties and rock mass properties, has demonstrated favorable outcomes. This implies that the equations are dependable in predicting TBM (Tunnel Boring Machine) performance and can be utilized in situations where machine parameters are lacking. The conclusion drawn was that intact rock properties serve as the primary input parameters for predicting TBM performance. However, relying solely on intact rock properties may be insufficient, as in cases of fragmented rock, they may not adequately reflect the strength of the rock mass. It is also important to note that using the prediction formula without machine parameters can lead to inaccurate results, since machine parameters are also volatile in different conditions and affect the performance of TBM in a complex way. Method can be used for more extensive analysis, but limitations such as unrealistic values when imputing data must be taken into account